A WAVELET BASED SIGNAL DENOISING AND HYDROCARBON PREDICTION ALGORITHM FOR A NEW MARINE CSEM ANTENNA DESIGN

• Main Author: NYAMASVISVA, TADIWA ELISHA
• Format: Thesis
• Language: English
• Published: 2017
• Subjects: Q Science (General)
• Online Access: http://utpedia.utp.edu.my/id/eprint/22034/1/Tadiwa%20Elisha%20Nyamasvisva%20%28PhD%29%20Thesis.pdf; http://utpedia.utp.edu.my/id/eprint/22034/

In Marine Controlled Source Electromagnetics (MCSEM) surveys, signals are congested with noisy elements such as airwaves, direct waves, magneto telluric waves, and reflected waves. Airwaves data are known to create ambiguities during interpretation of the presence of hydrocarbon (HC). So do other noisy elements, therefore there is a need to identify, quantify and eliminate these noises from the receiver data for better prediction of HC. This research work proposes a modified computational algorithm for decomposing and denoising the MCSEM data. Subsequently a mathematical model is derived for shallow water with deep hydrocarbon target environment. Currently MCSEM uses conventional horizontal electrical dipole (HED) antenna for deep water and shallow to deep HC exploration. This antenna lacks in directivity and focusing of the EM waves. The HED also propagates the EM waves equally in all directions which is a major problem in shallow water with air waves. In an attempt to explore deep HC in shallow waters, a new curved electric dipole (CED) antenna was designed with the capabilities of dispersing the airwaves by up to 77% in comparison with the conventional HED and able to enhance the down going signals for better resolution of up to 125%. Massive data sets are collected from extensive simulations of a geological model using the CED and HED antennae. A modified denoising algorithm was designed based on Johnstone and Donoho’s classical wavelet denoising protocol. The modified algorithm uses Symlet wavelet with soft threshold rule and Fixed Form Threshold (FFT) threshold technique, to decompose received signal Ex into single components comprising guided waves Gw, direct waves Dw, air waves Aw and magneto telluric wavesMT. Symlet 2 was taken as the base wavelet for decomposing MCSEM data on the basis that it filters 28% more residual data and takes on average 13% less computational time compared to other wavelets.